YaLM-100B: Pretrained language model with 100B parameters
1–10 of 666 posts
Re: YaLM-100B: Pretrained language model with 100B parameters
#2I hope one day it will be possible to run this kind of models at home.
Re: YaLM-100B: Pretrained language model with 100B parameters
#3I hope one day it will be possible to run this kind of models at home.
If you live in a datacenter, it already is!
Re: YaLM-100B: Pretrained language model with 100B parameters
#4I hope one day it will be possible to run this kind of models at home.
I was about to comment exactly the same thing. Stuff like this makes me feel so much behind because there's no way I can run this lol.
Re: YaLM-100B: Pretrained language model with 100B parameters
#5For those of us without 200GB of GPU RAM available... How possible is it to do inference loading it from SSD?
Would you have to scan through all 200GB of data once per character generated? That doesn't actually sound too painful - 1 minute per character seems kinda okay.
And I guess you can easily do lots of data parallelism, so you can get 1 minute per character on lots of inputs and outputs at the same time.
Re: YaLM-100B: Pretrained language model with 100B parameters
#6I have huge respect for developers at Yandex. It's kind of sad that achievements like these are tainted by the fact that they come from Russia (and I speak as a Ukrainian). I wonder if the permissive license is able to mitigate that.
Re: YaLM-100B: Pretrained language model with 100B parameters
#7I hope one day it will be possible to run this kind of models at home.
Well, it used to be impossible to render on anything not a mainframe in a reasonable time.
The day will come when we will be able to.
Re: YaLM-100B: Pretrained language model with 100B parameters
#8I have huge respect for developers at Yandex. It's kind of sad that achievements like these are tainted by the fact that they come from Russia (and I speak as a Ukrainian). I wonder if the permissive license is able to mitigate that.
The achievements aren’t in any way tainted by their nationality, citizenship, sex, sexual orientation, age, etc.
Re: YaLM-100B: Pretrained language model with 100B parameters
#9I have huge respect for developers at Yandex. It's kind of sad that achievements like these are tainted by the fact that they come from Russia (and I speak as a Ukrainian). I wonder if the permissive license is able to mitigate that.
Coming from Russia doesn't mean you agree with government policy. If you saw people get arrested as soon as they start protesting, what would you do?
Re: YaLM-100B: Pretrained language model with 100B parameters
#10I hope one day it will be possible to run this kind of models at home.
I was about to comment exactly the same thing. Stuff like this makes me feel so much behind because there's no way I can run this lol.
They hardware they mention can be rented from cloud providers. It’s just that it’s not very cheap.